Saumik Bhattacharya

Saumik Bhattacharya

Auto-generated profile, based on publicly available information

Saumik Bhattacharya

IIT Kharagpur

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 November 2020 in Applied Soft Computing
Verified
Journal Article
Closed access
SGSayantari Ghosh
Sayantari Ghosh
SBSaumik Bhattacharya
Saumik Bhattacharya

22 April 2021 in SN Computer Science
Verified
Journal Article
Closed access
PDF file
SGSayantari Ghosh
Sayantari Ghosh
SBSaumik Bhattacharya
Saumik Bhattacharya

10 January 2023 in Pattern Recognition
Verified
Journal Article
Closed access
PRPrasun Roy
Prasun Roy
SBSaumik Bhattacharya
Saumik Bhattacharya
SGSubhankar Ghosh
Subhankar Ghosh
UPUmapada Pal
Umapada Pal

3 January 2025 in IEEE Transactions on Artificial Intelligence
Verified
Journal Article
Closed access
SMSiladittya Manna
Siladittya Manna
SCSoumitri Chattopadhyay
Soumitri Chattopadhyay
RDRakesh Dey
Rakesh Dey
UPUmapada Pal
Umapada Pal
SBSaumik Bhattacharya
Saumik Bhattacharya
In contemporary self-supervised contrastive algorithms like SimCLR, MoCo, etc., the task of balancing attraction between two semantically similar samples and repulsion between two samples of different classes is primarily affected by the presence of hard negative samples. While the InfoNCE loss has been shown to impose penalties based on hardness, the temperature hyper-parameter is the key to regulating the penalties and the trade-off between uniformity and tolerance. In this work, we focus our attention on improving the performance of InfoNCE loss in self-supervised learning by proposing a novel cosine similarity dependent temperature scaling function to effectively optimize the distribution of the samples in the feature space. We also provide mathematical analyses to support the construction of such a dynamically scaled temperature function. Experimental evidence shows that the proposed framework outperforms the contrastive loss-based SSL algorithms. Our code is available at https://www.github.com/sadimanna/DySTreSS.

8 December 2017 in IEEE Transactions on Image Processing
Verified
Journal Article
Closed access
SBSaumik Bhattacharya
Saumik Bhattacharya
KVK. Subramanian Venkatesh
K. Subramanian Venkatesh
SGSumana Gupta
Sumana Gupta
We propose a novel technique for detection of visual saliency in dynamic video based on video decomposition. The decomposition obtains the sparse features in a particular orientation by exploiting the spatiotemporal discontinuities present in a video cube. A weighted sum of the sparse features along three orthogonal directions determines the salient regions in the video cubes. The weights computed using the frame correlation along three directions are based on the characteristic of Human Visual system (HVS) that identifies the sparsest feature as the most salient feature in a video. Unlike the existing methods, which detect the salient region as blob, the proposed approach detects the exact boundaries of salient region with minimum false detection. The experimental results confirm that the detected salient regions of a video closely resemble the salient regions detected by actual tracking of human eyes. The algorithm is tested on different types of video contents and compared with several state of the arts methods to establish the effectiveness of the proposed method.

4 December 2024
Verified
Book Chapter
Closed access
PDF file
ADAlloy Das
Alloy Das
SBSanket Biswas
Sanket Biswas
UPUmapada Pal
Umapada Pal
JLJosep Llados-Masllorens
Josep Llados-Masllorens
SBSaumik Bhattacharya
Saumik Bhattacharya